Satellite images show where tree canopy disappeared. Search your commune or a mountain and follow 25 years on the 3D terrain map. The map shows the place and year without establishing what caused the loss.
Loading the terrain…
Or scroll: the story in five steps.
County totals and the comparison with official statistics are below: counties and official figures.
2001–2025
A quarter of a century, on the map
Each orange patch is a place where the satellite saw the forest canopy disappear in some year. Yellow shows the year in the corner of the screen. Orange gathers the years before it.
In 2000, 7.9 million hectares had a dense canopy (at least 30% cover). On 476,000 hectares of them the canopy was gone by 2025: 6%, or about 19,000 hectares a year. The series shows no clear rise or fall.
2001–2025
Not the same everywhere
Suceava, Harghita and Maramureș hold 14% of the country's area with tree canopy in 2000 and 30% of the losses. In Suceava the canopy disappeared on 70,708 hectares, 15.2% of its area with canopy in 2000. The national figure is 6%.
Canopy loss is not deforestation. It includes legal cuts followed by replanting.
Colour by year
Patches climb the valleys
Here colour shows the year: from blue-violet in 2001 to yellow in 2025. On some valleys the patches line up in order, from the valley floor towards the ridge, the way a forest road advances.
The map does not identify roads or show whether cutting caused the canopy loss. It shows only that the places come in a sequence, year after year.
What is measured
What "canopy lost" means
A pixel covers about 0.05 hectares in Romania (28 × 19 metres). If in 2000 trees over five metres covered at least 30% of the pixel and a later disturbance removed the canopy, the dataset records the year of loss.
Turquoise marks where the dataset detected tree growth between 2000 and 2012. Growth after 2012 cannot be seen in these data.
Protected areas
Protected areas are not exempt
Natura 2000 sites cover 6.1 million hectares and 38% of the area with canopy in 2000. Canopy disappeared on 6.5% of that area inside the sites, against 5.7% outside them. In the national parks, 3.3%.
The data do not show how many losses came from cutting or whether those cuts were authorised.
Put your place on the map
Open the map, search a commune, a county or a protected area and choose the year at the bottom. For each place the panel shows the loss by year, against the national average.
How much, year by year?
The average was 18,400 hectares a year in 2001–2010 and 18,800 in 2013–2025. The linear trend's p-value is 0.5, above 0.05, so it is not statistically significant. The series is jagged, though, from 9,400 hectares in 2003 to 38,100 in 2007.
▮ annual loss, hectares (satellite)
━ three-year average
┅ regeneration cuts in high forest (INS)
━ clear-cuts (INS)
Why 2007 and 2012 jump
In 2007 the satellite saw 38,100 hectares, and in 2012 30,700. In 2007, 80% of the surplus over the neighbouring years sits in five counties: Suceava, Cluj, Bistrița-Năsăud, Harghita and Maramureș. So it is not a small error spread across the country, and we do not know what produced it: cutting, a storm, pests, or simply more and clearer images that year.
The dataset authors state that the years are not strictly comparable: the original map for 2001–2012 was made from a single run of the algorithm, and from 2011 onwards the data were recomputed, with Landsat 8 (from 2013) and Landsat 9 (from 2022) images. The years before 2011 were not recomputed. That is why the dashed blue line on the charts marks 2011.
Does it match the official statistics?
The short answer: in part. We put the satellite data next to the statistics of the National Institute of Statistics (INS) and of the National Forest Inventory. Three comparisons, each with its own check.
Size: the satellite sits between clear-cuts and regeneration cuts
In every year from 2001 to 2025, the loss seen from the satellite (9,400–38,100 hectares) is larger than the area of clear-cuts in INS statistics (2,437–5,480 hectares) and smaller than the area of regeneration cuts in high forest (46,000–79,000 hectares). Checked year by year: 25 out of 25. The order makes sense: the satellite also catches cuts that open the canopy over whole pixels without being clear-cuts, but it misses cuts that leave the canopy standing.
Place: it matches at county level
Counties: timber harvested, 2001–2024 (INS), and hectares of canopy lost. Logarithmic scale.Years: timber harvested per year nationally (INS) and hectares of canopy lost, 2001–2024.
Across the 41 counties with data, the ranking by hectares of canopy lost and the ranking by volume harvested agree well (rank correlation: 0.9). The area with trees may contribute to this correlation. When canopy loss is divided by canopy area in 2000 and harvested volume by the INS forest area in 2020, the correlation drops to 0.4.
Time: the years do not match
Year by year, the correlation between hectares seen from the satellite and the volume harvested nationally is only 0.1 (2001–2024). The INS volume rises slowly; the satellite series jumps. A 2020 study in Nature (Ceccherini and colleagues) found 0.39 for Romania between the Hansen data (version 1.6, 2004–2018) and officially recorded harvests; the correlation is not directly comparable with the one here, computed on version 1.13 and the years 2001–2024.
Timber: the satellite sees only part of what is cut
If every hectare of lost canopy had carried the Inventory's mean standing volume (340 cubic metres per hectare), the 453,000 hectares of 2001–2024 would amount to about 154 million cubic metres: a hypothetical standing-stock equivalent, not a measurement. Over the same period INS recorded 427 million cubic metres harvested. The gap cannot be broken down from these data.
Forest area: almost a million hectares apart
In 2000 the satellite finds 7.9 million hectares with canopy of at least 30% (with the 10% threshold used by FAO: 8.5). The National Forest Inventory counts 6.93 million hectares of forest in 2013–2018 and 7.03 in the third cycle. The gap of about 0.9 million hectares results from comparing different sources, periods and definitions (satellite in 2000, inventory in 2013–2025); these data do not show how much each contributes. The satellite counts any tall canopy, including orchards, shelterbelts, riverside thickets and tall scrub, while the inventory generally uses a threshold of more than 10% canopy cover (on land larger than 0.5 hectares, with trees able to reach 5 metres and at least 20 metres wide) and also includes some temporarily treeless areas expected to regenerate. These areas do not represent the same definition of forest.
The Forest Inventory counts 6.93 million hectares in 2013–2018 and 7.03 in the third cycle; with sampling errors of about 1% and method changes between cycles, the comparison does not establish a trend. Canopy lost is not forest lost: in many places a cut is followed by regeneration or planting.
By county
Hectares of canopy lost in 2001–2025, as a number and as a share of the area with canopy in 2000, and the volume harvested recorded by INS in 2001–2024. Sorted by hectares.
Counties, by hectares of canopy lost
County
Area with canopy in 2000, ha
Canopy lost 2001–2025, ha
Of the area with canopy in 2000
Timber harvested 2001–2024, thousand m³ (INS)
Suceava
464,660
70,708
15.2%
55,050
Harghita
326,383
44,389
13.6%
28,625
Maramureș
353,228
29,120
8.2%
14,334
Cluj
222,766
22,723
10.2%
6,287
Argeș
344,936
22,079
6.4%
14,906
Bacău
299,647
19,650
6.6%
21,149
Neamț
285,177
18,373
6.4%
27,735
Bihor
245,721
17,132
7%
11,222
Alba
274,071
16,338
6%
11,651
Mureș
258,428
16,125
6.2%
17,946
Bistrița-Năsăud
240,408
15,957
6.6%
13,625
Caraș-Severin
528,024
14,963
2.8%
17,115
Hunedoara
445,164
14,555
3.3%
13,426
Arad
269,310
13,848
5.1%
14,579
Sibiu
247,439
13,509
5.5%
11,290
Brașov
242,489
13,019
5.4%
15,764
Covasna
183,171
11,452
6.3%
12,668
Vâlcea
354,350
11,320
3.2%
11,221
Vrancea
202,236
10,881
5.4%
11,842
Gorj
307,204
8,684
2.8%
9,836
Timiș
136,600
6,619
4.8%
6,953
Buzău
176,547
6,409
3.6%
8,845
Dolj
67,005
5,052
7.5%
4,221
Mehedinți
173,330
4,854
2.8%
4,785
Satu Mare
84,361
4,840
5.7%
3,510
Prahova
190,108
4,614
2.4%
9,650
Tulcea
240,697
4,599
1.9%
4,261
Dâmbovița
148,308
4,549
3.1%
6,857
Iași
96,346
4,401
4.6%
6,663
Călărași
19,841
3,670
18.5%
3,143
Sălaj
136,312
3,341
2.5%
4,387
Ialomița
21,090
2,523
12%
2,456
Olt
49,271
2,416
4.9%
2,206
Botoșani
56,820
2,292
4%
3,705
Brăila
20,461
2,130
10.4%
1,846
Vaslui
75,805
1,871
2.5%
4,266
Teleorman
24,626
1,786
7.3%
1,456
Galați
25,927
1,658
6.4%
1,573
Giurgiu
35,913
1,645
4.6%
2,759
Constanța
26,881
1,400
5.2%
1,468
Ilfov
25,656
821
3.2%
1,740
București
905
16
1.8%
–
Communes with the most hectares
The places below have the largest area of canopy lost, not necessarily the highest rate. The figures show only the canopy seen from the satellite: they do not say who cut, or whether it was legal.
Communes with the most hectares of canopy lost
Place
Area with canopy in 2000, ha
Canopy lost 2001–2025, ha
Of the area with canopy in 2000
Măguri-Răcătău · Cluj
20,126
8,044
40%
Borșa · Maramureș
28,662
7,046
24.6%
Cârlibaba · Suceava
24,601
6,009
24.4%
Budureasa · Bihor
24,176
5,859
24.2%
Broșteni · Suceava
37,865
5,430
14.3%
Vișeu de Sus · Maramureș
36,639
4,639
12.7%
Beliș · Cluj
16,540
4,524
27.4%
Toplița · Harghita
22,149
4,354
19.7%
Nucșoara · Argeș
33,152
3,701
11.2%
Tulgheș · Harghita
19,562
3,544
18.1%
Poiana Stampei · Suceava
17,806
3,535
19.9%
Ibănești · Mureș
27,175
3,261
12%
Jina · Sibiu
23,968
3,233
13.5%
Lunca Bradului · Mureș
28,134
3,216
11.4%
Asău · Bacău
27,471
3,211
11.7%
Poienile de sub Munte · Maramureș
21,431
3,099
14.5%
Rucăr · Argeș
21,249
3,063
14.4%
Moldovița · Suceava
24,760
2,984
12.1%
Bilbor · Harghita
16,507
2,945
17.8%
Voineasa · Vâlcea
35,794
2,857
8%
Nistorești · Vrancea
21,721
2,850
13.1%
Gura Teghii · Buzău
41,398
2,829
6.8%
Șugag · Alba
19,081
2,820
14.8%
Gheorgheni · Harghita
13,275
2,716
20.5%
Agăș · Bacău
14,415
2,646
18.4%
What the map cannot know
Cause of loss and permits
A marked pixel shows canopy that disappeared without establishing the cause. If a cut took place, the data do not say who made it or whether it was authorised. Figures on illegal logging in Romania are contested and we do not use them as a benchmark.
What was there, what grew, how old it was
We cannot see species, age or condition of the forest, or whether the place was replanted or regrew by itself.
Cuts that leave the canopy standing
Thinnings and selective cuts can remove a lot of timber without emptying the pixel.
Growth after 2012
The gain layer covers only 2000–2012, so the map cannot show net change in forest. The data layer indicates 154,000 hectares of gain in that period.
Image errors
Clouds, shadows on steep slopes and sensor changes can move a patch from one year to another or create false patches. Annual figures change between versions of the data.
Edges
Protected-area outlines are simplified to about 90 metres, and commune boundaries to about 30 metres. At the edges a pixel can be counted in the neighbouring place.
How we calculated it and where the data come from
Canopy loss. From the Global Forest Change dataset (Hansen, University of Maryland, Google, USGS and NASA), version 1.13, for 2000–2025, from Landsat images on a 0.00025-degree grid (about 28 × 19 metres in Romania). We used one field: the year of loss (1–25, meaning 2001–2025), on pixels with at least 30% canopy cover in the year 2000. Each pixel's area is computed by latitude (about 0.054 hectares at 46° north), not at the equator value. With the 10% threshold the total is 480,000 hectares (5.6%).
External check. For 2001–2024 our total is 453,000 hectares, 5.7% of the 2000 canopy. ECSR reports 453,000 hectares, 5.7%, attributing the figure to Global Forest Watch; we could not verify it directly.
Terrain. Copernicus GLO-30 (European Space Agency), 30-metre resolution, reduced to about 100 metres on the page. The vertical exaggeration varies with the distance from the camera.
Boundaries and zones. Communes and counties: geo-spatial.org, from ANCPI data (CC BY-SA 4.0). Protected areas: Natura 2000 sites, 2024 version (European Environment Agency, CC BY 4.0), and national and natural parks (National Agency for Protected Natural Areas, through the European Environment Agency, NatDA v23, CC BY 4.0). The short names in the search show which outline we use: "Făgăraș", for example, is the Natura 2000 site Munții Făgăraș; "Obcinele Bucovinei" is the site of that name, not all of Bucovina.
Official statistics. INS, TEMPO-Online: area cut (AGR305A), harvested volume by county (AGR306A), forest fund (AGR301A). National Forest Inventory, INCDS "Marin Drăcea": cycle 2 (2013–2018) and cycle 3.
Sources
Hansen, M. C. and colleagues, 2013, "High-Resolution Global Maps of 21st-Century Forest Cover Change", Science 342: 850–853. Data: glad.earthengine.app/view/global-forest-change, version GFC-2025-v1.13. Source: Hansen/UMD/Google/USGS/NASA. Licence: CC BY 4.0. We aggregated the pixels by commune and county and drew them on the terrain.
National Forest Inventory, INCDS "Marin Drăcea": roifn.ro.
Ceccherini, G. and colleagues, 2020, "Abrupt increase in harvested forest area over Europe after 2015", Nature 583: 72–77: nature.com/articles/s41586-020-2438-y.
A total of 453,000 hectares for 2001–2024, attributed by ECSR to Global Forest Watch: ecsr.ro.
Frequently asked questions
How much forest has been cut in Romania since 2001?
These data cannot tell us how many hectares were cut. Satellites see canopy lost on 476,000 hectares in 2001–2025 (6% of the area with canopy in 2000), about 19,000 hectares a year. The figure includes legal harvest, wind, fire and pests, and does not subtract areas where trees grew back.
Which county lost the most tree canopy?
Suceava: 70,708 hectares, 15.2% of its area with canopy in 2000. Harghita (44,389 hectares) and Maramureș (29,120 hectares) follow.
Does canopy loss mean illegal logging?
No. The map shows where and when canopy disappeared, but not whether cutting took place, who did it or whether it was authorised.
Share
Trees in 2000
Canopy lost
Canopy lost in the chosen year
Below 30% canopy cover in 2000
20012025Colour shows the year the canopy disappeared.